InfluxDB Schema Design

2
votes
  • I have 3000 Sensors 1500 of temperature 1000 of humidity and 500 of pressure
  • Each of these sensors are mapped to a location , i.e I have 3000 locations
  • Location can be like Building/Block/Floor/Room/LocA and Building/Block/Floor/Room/LocB (this being the parent Building/Block/Floor/Room )
  • data is read from the sensors every 15 seconds

What should be my schema for InfluxDB ?

Here are my thoughts

I thought I will have these as measurements

  1. temperature
  2. pressure
  3. humidity

And the Tags as Location | sensorID (3000 tags for all the sensors which makes my cardinality 3000 )

Location | sensorID 
---------|---------
location A| sensor1
location B| sensor2

so I will have 3000 series, data stored every 15 seconds Is this scalable? Do you see any issues with this schema design

1
influxdb
seems fine, but you might want to add additional tag for Building/Block/Floor/Room if you will ever want to group or compare metrics per "Block". In terms of scaling will you have more series in the future ? - Tomasz Gawlik
@TomaszGawlik Yes I expect it to grow based on the client , how does this look for the current load This is a pilot project, later I have to deliver solution for a client with 8000 sensors - Purna
Do you see any issue with scalability and I need to backup this data periodically - Purna

1 Answers

1
votes

Seems fine if the system is more than 8GB RAM in large scale.Please refere the following doc from influxdb https://docs.influxdata.com/influxdb/v1.7/guides/hardware_sizing/